Vertical creator short
Use 9:16, a tighter crop, and a more casual creator expression.
Avatar prompt
Vidu AI Talking Avatar Prompt Recipe
Use this talking avatar recipe to build a Vidu AI AI video prompt with subject, action, camera, lighting, style, and constraints.
Copy-ready prompt
Subject: a professional presenter in a simple dark blazer. Action: speaking calmly with subtle head movement and natural blinking. Scene: a clean studio background with a soft depth-of-field backdrop. Motion: locked medium close-up at eye level. Look: talking avatar video, realistic presenter, clear facial consistency. Lighting: soft key light with gentle fill and natural skin tone. Format: 8 seconds, 9:16. Negative: lip-sync mismatch, exaggerated mouth movement, drifting identity, frozen eyes, extra teeth.
Advertisement
Prompt notes
- This version separates subject, action, scene, motion, look, and negative constraints.
- subject: a professional presenter in a simple dark blazer
- action: speaking calmly with subtle head movement and natural blinking
- setting: a clean studio background with a soft depth-of-field backdrop
- camera: locked medium close-up at eye level
- lighting: soft key light with gentle fill and natural skin tone
- style: talking avatar video, realistic presenter, clear facial consistency
- negative: lip-sync mismatch, exaggerated mouth movement, drifting identity, frozen eyes, extra teeth
- aspectRatio: 9:16
- duration: 8 seconds
- Big hand gestures can distract from facial consistency in a short talking-head clip.
- Complex backgrounds make presenter identity and mouth motion harder to evaluate.
Use this when
- You need a presenter, spokesperson, explainer, or creator-style talking-head video draft.
- The face, gaze, mouth motion, and posture need to stay stable across the clip.
- The prompt should support a clean avatar shot without distracting background motion.
How to adapt each field
subject
Define the presenter with stable visible traits such as clothing, posture, age range, and expression.
action
Use calm speaking motion, subtle gestures, and natural blinking instead of broad dramatic performance.
setting
Keep the background simple so the viewer focuses on the face and message.
camera
Locked medium close-up framing helps facial consistency and reduces body distortion.
lighting
Soft lighting and natural skin tone language are important for believable talking-head clips.
style
Use presenter, spokesperson, explainer, or creator terms depending on the intended channel.
negative
Talking-avatar prompts need explicit constraints for mouth motion, eyes, teeth, and identity drift.
Drafting workflow
- Set the presenter identity and camera framing before adding performance notes.
- Keep the first draft to one speaking beat with minimal hand movement.
- Check mouth, eyes, face shape, and shoulders for frame-to-frame stability.
- Add script timing, gesture, or branded background details only after the face stays stable.
Useful variations
Corporate explainer
Use neutral studio lighting, formal clothing, and minimal gestures.
Product spokesperson
Add a product on a side table or subtle brand-colored backdrop while keeping face motion simple.
Talking Avatar prompt FAQ
Why use a locked camera for talking-avatar prompts?
A locked camera reduces competing motion. That makes it easier to judge face, gaze, mouth movement, and identity consistency.
Should I include the full script in the prompt?
For a visual prompt, keep the script short or describe the speaking beat. Exact narration and lip sync usually depend on the model interface or editing workflow.
What are the most common talking-avatar artifacts?
Mouth shapes, teeth, eyes, face identity, and shoulder motion are common weak spots, so the negative prompt should name them directly.
Related prompt pages
Independent prompt drafting aid. Verify final prompts inside the current model interface.